Agent Tools Orchestration Leaks More: Dataset, Benchmark, and Mitigation

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Qiao, Yuxuan, Liu, Dongqin, Yang, Hongchang, Zhou, Wei, Hu, Songlin
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866916071199473664
author Qiao, Yuxuan
Liu, Dongqin
Yang, Hongchang
Zhou, Wei
Hu, Songlin
author_facet Qiao, Yuxuan
Liu, Dongqin
Yang, Hongchang
Zhou, Wei
Hu, Songlin
contents LLM-based agents increasingly use multiple external tools to complete complex tasks. We study Tools Orchestration Privacy Risk (TOP-R): an agent may combine individually non-sensitive tool returns and disclose an unintended sensitive conclusion. We formalize TOP-R with three conditions: conclusion sensitivity, single-source non-inferability, and compositional inferability. We introduce LRSE (Library-Grounded Reverse-Inference Seed Expansion), a four-library reverse-construction pipeline grounded in privacy norms, reasoning chains, tool schemas, and task scenarios, and use it to build TOP-Bench, a 1,000-instance benchmark. The benchmark evaluates final-response semantic disclosure under a controlled two-stage tool-use protocol. Across six LLM agents, task completion remains high, but the average leakage rate reaches 88.6 percent, yielding an H-score of only 20.4. Two prompt-only safeguards improve H-score by about 2.7 points on the main benchmark. We further propose TOP-Align, an SFT+DPO post-training method for safer task completion boundaries. On a separate post-training evaluation split, TOP-Align improves H-score by 16.2 points over the corresponding base model, compared with a 4.9-point average gain from prompt-only mitigation on the same split. These results show that TOP-R requires mitigation beyond prompting alone.
format Preprint
id arxiv_https___arxiv_org_abs_2512_16310
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Agent Tools Orchestration Leaks More: Dataset, Benchmark, and Mitigation
Qiao, Yuxuan
Liu, Dongqin
Yang, Hongchang
Zhou, Wei
Hu, Songlin
Cryptography and Security
Artificial Intelligence
Computation and Language
LLM-based agents increasingly use multiple external tools to complete complex tasks. We study Tools Orchestration Privacy Risk (TOP-R): an agent may combine individually non-sensitive tool returns and disclose an unintended sensitive conclusion. We formalize TOP-R with three conditions: conclusion sensitivity, single-source non-inferability, and compositional inferability. We introduce LRSE (Library-Grounded Reverse-Inference Seed Expansion), a four-library reverse-construction pipeline grounded in privacy norms, reasoning chains, tool schemas, and task scenarios, and use it to build TOP-Bench, a 1,000-instance benchmark. The benchmark evaluates final-response semantic disclosure under a controlled two-stage tool-use protocol. Across six LLM agents, task completion remains high, but the average leakage rate reaches 88.6 percent, yielding an H-score of only 20.4. Two prompt-only safeguards improve H-score by about 2.7 points on the main benchmark. We further propose TOP-Align, an SFT+DPO post-training method for safer task completion boundaries. On a separate post-training evaluation split, TOP-Align improves H-score by 16.2 points over the corresponding base model, compared with a 4.9-point average gain from prompt-only mitigation on the same split. These results show that TOP-R requires mitigation beyond prompting alone.
title Agent Tools Orchestration Leaks More: Dataset, Benchmark, and Mitigation
topic Cryptography and Security
Artificial Intelligence
Computation and Language
url https://arxiv.org/abs/2512.16310